Amlexanox Downregulates S100A6 to Sensitize KMT2A/AFF1-Positive Acute Lymphoblastic Leukemia to TNFα Treatment

The most prevalent mixed-lineage leukemia (MLL) rearrangement in acute lymphoblastic leukemia (ALL) generates the KMT2A/AFF1 (MLL/AF4) fusion gene due to the t(4;11)(q21;q23) chromosomal translocation. ALL with t(4;11)(q21;q23) has a bimodal age distribution with a major peak of incidence in early infancy and accounts for over 50% of ALL cases in infants < 6 months old, 10% to 20% in older infants, 2% in children, and up to 7% in adults (1). Despite recent improvements in the overall treatment outcome for ALL patients, including allogeneic hematopoietic stem cell transplantation (allo-HSCT), KMT2A/AFF1-positive ALL is still associated with a poor prognosis (2). The complete remission (CR) rate in children is as high as 88%, but the median overall survival (OS) is only 10 months, indicating an extremely poor prognosis. In adult patients with ALL, the CR rate is 75%, but the prognosis is also poor, with a median OS of 7 months (1).

The poor prognosis of KMT2A/AFF1-positive ALL has been suggested to be due to resistance to TNFα, which is the factor involved in the graft-versus-leukemia (GVL) effect, or tumor immunity by upregulation of S100A6 expression followed by interference with the p53–caspase pathway (3). S100A6 is a 10.5-kDa Ca²⁺-binding protein belonging to the S100 protein family, which has been reported to interact with and alter the conformation of p53 (4–7). Upregulation of S100A6 expression in KMT2A/AFF1-positive ALL inhibits p53 acetylation followed by inhibition of caspase apoptotic pathway upregulation in the presence of TNFα (8).

Here, we focused on amlexanox (2-amino-7-isopropyl-5-oxo-5H-chromeno[2,3-b] pyridine-3-carboxylic acid), a common anti-allergic drug, which targets S100A6. Amlexanox was reported to inhibit the translocation pathway of endogenous S100A6 in endothelial cells (9). Amlexanox has also been reported to bind to S100A13, which is another member of the S100 protein family (10). S100A13 and acidic fibroblast growth factor (FGF1) are involved in a wide range of important biological processes, including angiogenesis, cell differentiation, neurogenesis, and tumor growth. Generally, the biological function of FGF1 is to recognize a specific tyrosine kinase on the cell surface and initiate the cell signal transduction cascade. Amlexanox binds S100A13 and FGF1 and inhibits the heat shock–induced release of S100A13 and FGF1 (10). Amlexanox has been used in a number of recent metabolic studies. Reilly and colleagues reported that amlexanox acted as an inhibitor of IKKϵ and TBK1 (11). They showed that treatment with amlexanox resulted in reduced inflammation, marked improvement in insulin sensitivity, and reduced hepatosteatosis in mice with genetic or dietary obesity (11). The present study was performed to examine the effects of amlexanox in KMT2A/AFF1-positive ALL.

The KMT2A/AFF1-positive human ALL cell line RS4;11 was purchased from ATCC. The KMT2A/AFF1-positive ALL cell line SEM was purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ). The cell lines were obtained in 3 years and used within 6 months after receipt or resuscitation. To check mycoplasma we used a PCR-based method, indirect staining and an agar and broth culture. SEM, SEM transduced with a lentiviral vector expressing luciferase (SEM-Luc), and RS4;11 cells were cultivated in RPMI 1640 (Sigma-Aldrich) supplemented with 10% FBS (PAN Biowest) at 37°C under 5% CO₂.

SEM and RS4;11 were seeded in 6-well plates (2 × 10⁵ cells/well) and incubated in vitro with TNFα (0 and 10 ng/mL; Wako) or Human peripheral blood mononuclear cells (PBMC; 2 × 10⁵ cells/well) and amlexanox (Tokyo Chemical Industry; 0, 10, and 100 μg/mL) for 48 hours before counting cells to examine the effects of TNFα and amlexanox on leukemia cells.

Western blotting analysis was performed as described previously (3). SEM and RS4;11 cells were incubated for 48 hours and collected for Western blotting analysis. Equal aliquots of lysates from cell lines or homogenized mouse spleen were subjected to 10% SDS-PAGE, transferred onto polyvinylidene difluoride membranes, and immunoblotted with the following primary antibodies (Abs): anti-S100A6 (calcyclin) Ab (Santa Cruz Biotechnology), anti–caspase-3, anti-cleaved caspase-3 Ab (Cell Signaling Technology), anti-p53 (Santa Cruz Biotechnology), anti–acetyl-p53 (Millipore), and anti–β-actin Ab (Millipore). Can Get Signal (Toyobo) was used to promote the reaction between primary Ab and antigen (Ag).

KMT2A/AFF1 transgenic (Tg) mice, which show CD45R/B220⁺ leukemia by 12 months of age, at which time lymphoma cells will have infiltrated the bone marrow (BM) and spleen, were established previously (12). This KMT2A/AFF1 Tg mouse model has an essentially normal immune system. We divided 10 male KMT2A/AFF1 Tg mice at the age of 4 months into two groups: the amlexanox group (n = 5) fed a diet containing amlexanox (0.02%) for 10 months, and the control group (n = 5). All KMT2A/AFF1 Tg mice were killed at the age of 14 months for histopathological and Western blotting analysis. All animal experiments were performed in accordance with the regulations established by the Ethical Committee of Nippon Medical School and were approved by the Animal Care and Committee of Nippon Medical School (Approval number: 28-026).

For histopathological analysis, mice were killed and tissues of interest were fixed in 4% paraformaldehyde. Hematoxylin and eosin (H&E) staining was performed using standard protocols. For immunopathological analysis, the slides were fixed for 30 minutes in 4% paraformaldehyde, microwaved for 5 minutes in Vector Antigen Unmasking Solution (pH 6.0; Vector Laboratories) for Ag retrieval, rinsed twice for 10 minutes each time in PBS, and incubated for 15 minutes in 3% hydrogen peroxide in methanol to quench endogenous peroxidase activity. For CD45R/B220 staining, the pretreated slides were incubated for 60 minutes with anti-CD45R/B220 (BD Pharmingen) primary Abs. To visualize anti-CD45R/B220 Ab binding, the slides were incubated for 30 minutes with FITC-conjugated anti-rat IgG (Jackson ImmunoResearch Laboratories). Nuclei were counterstained with Mayer's hematoxylin. The photomicrographs shown in the figures were taken with a SPOT Insight digital camera controlled using SPOT Advanced Version 4.0.9 software (Diagnostic Instruments) with a Nikon Eclipse 80i microscope equipped with Nikon Plan 2×/0.08 NA and Plan 40×/0.75 NA objectives. Comparison of CD45R/B220-positive tumor cells between the control group and amlexanox group was performed by counting the numbers of CD45R/B220-positive cells/mm².

Human PBMCs were obtained by separation of heparinized blood from a healthy donor on a Ficoll–Histopaque gradient. The PBMCs were washed twice and resuspended in RPMI 1640 supplemented with 25 mmol/L HEPES buffer, 2 mm glutamine, 100 U/mL penicillin, 100 pg/mL streptomycin, and 10% heat-inactivated FCS (designated as FCS-RPMI).

For in vivo analysis, 5 × 10⁵/body of SEM-Luc cells were injected intraperitoneally (i.p.) into three groups of 10 nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice. Each group of 10 mice was divided into two groups: the amlexanox + PBMC group (n = 5) fed a diet containing amlexanox (0.02%), and the PBMC group (n = 5) fed a diet without amlexanox. Feeding with each diet was started when SEM-Luc cells were injected and supplied consistently until mice died. Mice in each group were injected i.p. with 4 × 10⁷/body of human PBMCs every 2 weeks. Non–PBMC-injected mice fed a diet without amlexanox were used as controls. In addition to overall survival (OS) rate, tumor growth after injection of human PBMCs was assessed using an in vivo imaging system (IVIS) as described previously (3).

The results of cell growth were analyzed by the Student t test, assuming unequal variances and two-tailed distributions. Data are shown as the means ± standard deviation of at least three samples. For survival analyses, event time distributions were estimated using the method of Kaplan and Meier, and differences in survival rates were compared using the log-rank test. In all analyses, P < 0.05 was taken to indicate statistical significance.

First, to analyze the effects of amlexanox on KMT2A/AFF1-positive ALL cell lines, SEM and RS4;11 were seeded in 6-well plates (2 × 10⁵ cells/well) and incubated in vitro with TN-α (0 and 10 ng/mL; Wako) and amlexanox (Tokyo Chemical Industry; 0, 10, and 100 μg/mL). After 48 hours of incubation, viable cells were counted using Trypan blue exclusion. KMT2A/AFF1-positive ALL cells (SEM or RS4;11) showed no significant growth inhibition by 10 ng/mL of TNFα in the absence or presence of 10 μg/mL of amlexanox (Fig. 1A and B). However, both cell lines showed significant growth inhibition by 10 ng/mL of TNFα in the presence of 100 μg/mL of amlexanox (P = 0.0085 for SEM, P = 0.0196 for RS4;11; Fig. 1A and B). Next, to examine direct evidence that amlexanox increases cell-mediated anti-leukemia effects, SEM and RS4;11 cells were incubated with PBMCs (2 × 10⁵ cells/well) instead of 10 ng/mL of TNFα and viable cells were counted. SEM cells showed significant growth inhibition on incubation with PBMCs in the presence of 10 or 100 μg/mL of amlexanox (P = 0.0136 or 0.0019; Fig. 2A). RS4:11 cells also showed significant growth inhibition on incubation with PBMCs in the presence of 100 μg/mL of amlexanox (P = 0.0125; Fig. 2B). These results indicated that amlexanox increases both TNFα and cell-mediated antileukemia effects.

To analyze the molecular mechanism underlying the effect of amlexanox on the KMT2A/AFF1-positive human ALL cells, SEM and RS4;11 were incubated for 48 hours, and collected for Western blotting analysis. As shown in Fig. 3, S100A6 was activated in the presence of 10 ng/mL TNFα, and activated S100A6 was decreased and both acetyl-p53/p53 ratio and cleaved caspase-3/caspase-3 ratio were increased in cells treated with 100 μg/mL of amlexanox in the presence of 10 ng/mL of TNFα in the KMT2A/AFF1-positive human ALL cell line (Fig. 3A and B).

To examine whether amlexanox in combination with tumor immunity suppressed the development of MLL/AF4-positive ALL, we used KMT2A/AFF1-positive Tg mice in which the immune system is normal (12). In vivo analysis was performed using KMT2A/AFF1-positive Tg mice, which we established previously and show CD45R/B220⁺ pro–B-cell leukemia by 12 months of age (12). Although pro–B cell-leukemia infiltration was observed in all five mice in the control group, no leukemia occurred in four of five mice in the amlexanox group at the age of 14 months. Figure 4A (top) shows a comparison of the results of H&E staining between one mouse from the amlexanox group (amlexanox#1) and one mouse from the control group (control#1). Although the normal structure of the spleen was disrupted in control#1, no such effect was seen in the spleen of amlexanox#1. Although BM of control#1 was hypercellular with tumor cells, that of amlexanox#1 was normocellular with no tumor invasion. Figure 4A (bottom) shows a comparison of the results of immunohistopathologic analysis between one mouse from the amlexanox group (amlexanox#1) and one mouse from the control group (control#1). Although the spleen and BM of control#1 showed CD45R/B220⁺ leukemia infiltration, no leukemia was observed in amlexanox#1.

Figure 4B shows a comparison of CD45/B220⁺ leukemia mass/mm² between the control group (n = 5) and amlexanox group (n = 5). Both the spleen and BM of the amlexanox group had significantly less CD45/B220⁺ leukemia cell invasion in comparison with the control group (both P < 0.001).

To examine the mechanism underlying the suppression of leukemia by amlexanox in KMT2A/AFF1 Tg mice, we performed Western blotting analysis of the lysates from the spleens of KMT2A/AFF1 Tg mice in the amlexanox group and compared the results with those for KMT2A/AFF1 Tg mice in the control group. Western blotting analysis indicated the inhibition of S100A6 and upregulation of p53 acetylation as well as cleaved caspase-3 levels in the amlexanox group in comparison with the control group (Fig. 5A).

Reilly and colleagues reported that amlexanox reduced obesity in mice by inhibition of IKKϵ and TBK1 (11), and this may be a problem when considering use of amlexanox in human infants. Therefore, we monitored body weight of mice treated with amlexanox. As shown in Fig. 5B, KMT2A/AFF1 Tg mice in the amlexanox group had significantly lower weight at the age of 14 months than those in the control group (Fig. 5C; P < 0.001).

To examine whether amlexanox in combination with GVL effect could suppress the development of KMT2A/AFF1-positive ALL, we used PBMC-NOD/SCID mice transplanted with KMT2A/AFF1-positive ALL. As shown in Fig. 6A, PBMC-NOD/SCID mice transplanted with SEM-Luc in the amlexanox + PBMC group showed significantly longer survival than those in the control group (P = 0.011). The PBMC group did not show significantly longer survival than those in the control group (P = 0.93).

Figure 6B and C show comparisons of the volumes of SEM-Luc between the amlexanox + PBMC group and control group. On day 84 following transplantation, the volume of SEM-Luc was significantly lower in the amlexanox group than the control group (P = 0.003). There was no significant difference in SEM-Luc between the PBMC group and control group (P = 0.120).

The results presented here suggested that amlexanox inhibits the resistance of KMT2A/AFF1 (MLL/AF4)-positive ALL to TNFα by downregulating S100A6 expression. The results of this study indicated that amlexanox blocks upregulation of S100A6 expression followed by interference with the p53–caspase pathway in KMT2A/AFF1-positive ALL both in vitro and in vivo. The hypothetical mechanism underlying these effects is shown in Fig. 7. With respect to the genomic mechanism of KMT2A/AFF1, KMT2A is an H3K4 histone methyltransferase, and leukemia transformation by KMT2A fusion requires the H3 lysine 79 (H3K79) methyltransferase DOT1L, which is recruited to the KMT2A fusion transcriptional complex. Suppression of DOT1L inhibited expression of AFF1 target genes, indicating that H3K79 methylation is a distinguishing feature of KMT2A/AFF1-positive ALL and plays a key role in maintenance of AFF1-driven gene expression (13). However, our therapeutic target of KMT2A/AFF1-positive ALL is not this genomic mechanism but the resistance of KMT2A/AFF1 (MLL/AF4)-positive ALL to TNFα by downregulating S100A6 expression.

We previously reported that the poor prognosis of KMT2A/AFF1-positive ALL may be caused by resistance to TNFα, which is the factor involved in the GVL effect, or tumor immunity by upregulation of S100A6 expression followed by interference with the p53–caspase apoptotic pathway (3). Amlexanox, known as an anti-allergic drug, may exploit the GVL effect or tumor immunity in KMT2A/AFF1-positive ALL because it inhibits the resistance of KMT2A/AFF1-positive ALL to TNFα. In vivo analysis using KMT2A/AFF1-positive Tg mice showed that amlexanox in combination with tumor immunity may suppress the development of MLL/AF4-positive ALL. Therefore, using amlexanox after CR may be effective to suppress relapse of KMT2A/AFF1-positive ALL.

The GVL effect on KMT2A/AFF1-positive ALL may also be induced by amlexanox in PBMC-NOD/SCID mice. Specifically, allogeneic hematopoietic stem cell transplantation (allo-HSCT) is expected to show beneficial effects in combination with amlexanox despite its lack of efficacy in KMT2A/AFF1-positive ALL patients reported to date (2). Interestingly, Spijkers-Hagelstein and colleagues reported that overexpression of the S100 protein family members S100A8 and S100A9 in KMT2A/AFF1-positive ALL was associated with failure to induce free-cytosolic Ca²⁺ and prednisolone resistance (14). These observations taken together with our previous study suggest that high levels of S100 protein expression by KMT2A/AFF1-positive ALL may be the main factors involved in therapy resistance and poor prognosis of KMT2A/AFF1-positive ALL. Amlexanox was also reported to bind to another member of the S100 protein family, S100A13 (10). Although there have been no reports regarding the relations between amlexanox and S100A8 and S100A9, it is possible that this drug affects these proteins.

Another major advantage of the use of amlexanox is that it is already a commonly used anti-allergy drug and its cost is not high in contrast to almost all other molecular target drugs. However, there is a problem when considering use of amlexanox as treatment for human infants. As reported by Reilly and colleagues, amlexanox prevents weight gain by inhibition of IKK-ϵ and TBK1 (11). Our KMT2A/AFF1 Tg mice fed amlexanox were also significantly smaller than those fed the control diet (P < 0.001; Fig. 5B and C). This is an important problem for the treatment of infants, although it would not be an issue in adults. Despite these problems, treatment targeting S100A6 by amlexanox may increase tumor immunity and the effects of hematopoietic stem cell transplantation against KMT2A/AFF1-positive leukemia, thus making it useful for treatment of KMT2A/AFF1-positive leukemia.

No potential conflicts of interest were disclosed.

Conception and design: H. Tamai, H. Yamaguchi K. Inokuchi

Development of methodology: H. Tamai, H. Yamaguchi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Tamai, H. Yamaguchi, K. Inokuchi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Tamai, H. Yamaguchi, M. Takatori, T. Kitano

Writing, review, and/or revision of the manuscript: H. Tamai, H. Yamaguchi, K. Miyake, S. Yamanaka, K. Fukunaga, K. Nakayama, K. Inokuchi

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Tamai, H. Yamaguchi

Study supervision: H. Tamai, H. Yamaguchi, K. Miyake, K. Inokuchi

The authors thank Dr. R. Van Etten of Harvard Medical School for the kind gift of the pMSCVneop230 BCR/ABL plasmid for establishment of our KMT2A/AFF1 Tg mice.

H. Tamai was supported in part by grant 40465349 from the Ministry of Health and Welfare of Japan and the Ministry of Education, Science, and Culture of Japan. H. Tamai was also supported by grants from SENSHIN Medical Research Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.